There is strong evidence that signs of autism spectrum disorders (ASD) are present and detectable late in the first year and early in the second year of life. Parents first identify concerns about development at a mean age of 14 months. Despite this, the mean age of first diagnosis of ASD in the United States is between 41/2 and 51/2 years (CDC, 2012). Early intensive intervention can be highly effective for young children with ASD, but is reserved for children with a formal diagnosis, making early identification imperative. A measure that could identify ASD risk before the full set of symptoms has emerged offers the opportunity for intervention when the brain has the greatest plasticity. The promise of earlier identification remains hampered, however, by measurement issues. Although ASD symptoms can be reliably reported by parents in school-age children, individual variation is much more difficult for parents to assess in children under the age of 3 years. We propose to develop a new video-based method of assessing early development. Taking advantage of thousands of hours of already collected and coded video of infants at risk for ASD, we propose to create a video library depicting a continuum of social, communication, and repetitive behaviors, from very typical to clearly delayed development. In collaboration with a company that develops software for families of children with autism, we will create a measure, implemented through a secure website, with which parents can rate similarity to their infant's behaviors and identify video that best represents their child. This measure will be of low burden to families, as it is brief and can be administered in a secure manner via internet. The first year of funding will be used to develop this new measure (Aim 1), called the Video-Referenced Infant Rating System for Autism (VIRSA), which will extend from 6 to 18 months of age. After an initial validation process, the VIRSA will be administered three times (6, 12, and 18 months) to parents of 90 infants at familial risk for ASD and 45 infants with no family history of ASD (Aim 2) The infants will then be followed to 36 months of age, when diagnostic outcome evaluations will be performed. We will compare the psychometric properties and predictive validity of the VIRSA to other measures and identify best cutoffs for identification. We will examine agreement between parent- and expert-ratings to determine the validity of the measure in parents with and without experience of ASD. The ultimate goal of the project is to develop a low-cost, low-burden measure that can detect ASD risk in infancy. Such a measure has significant public health applications, including earlier diagnosis, longitudinal tracking of development to identify childre requiring evaluation, screening of large community-based samples, and monitoring of individual intervention progress.